An image filing system is used with documents each composed of a plurality of pages, each page bearing image information. Such system is capable of allowing the operator access to any desired document by making a search for a single keyword or two or more keywords assigned to a particular document if the predetermined document name is unknown by the operator. Whether or not the documents listed on retrieval of the keyword or keywords includes the correct document the operator intends to view cannot be determined by the operator, however, unless the operator actually views each of the pages of the document on display. Is the image data is stored in a compressed form in the system, the image data retrieved must be processed to restore or extend the image data to its original form before the image represented by the data is displayed. A disproportionately large amount of time is required for such processing of the retrieved image data. This time is in addition to the time and labor which the operator is required to expend for reviewing each of the pages of the document.
These problems are ingeniously solved in an image filing system proposed in Japanese Provisional Patent Publications (Kokai) No. 61-224557 and No. 61-224558. In the image filing system therein disclosed, a "characterizing" image area of a document which represents a prominent feature of the document is designated at the time of registration of the document in the system. Consequently, only this characterizing image area of the document is displayed during successive inspection of the documents listed as a result of a keyword.
When there is an instruction signal entered to reference the documents listed as a result of a keyword search in known image filing systems of this type, it is required that a search be further made of the characterizing image area of each of the documents. This characterizing image area data is stored in compressed form. The data representative of the particular image area of the document is processed to restore the original state of the data before the image in the image area is to be displayed. An intricate procedure is required, however, for selectively extracting the data contained in the characterizing image area of the document and, in addition, there is a limitation in reducing the period of time required for such processing of the data.